Spin-state transition, magnetism and local crystal structure in Eu1-xCaxCoO3-δδ
A.N. Vasiliev1, T.M. Vasilchikova1, O.S. Volkova1, A.A. Kamenev1, A.R. Kaul1, T.G. Kuzmova1, D.M. Tsymbarenko1, K.A. Lomachenko2, A.V. Soldatov2, S.V. Streltsov3,4, J.-Y. Lin5, C.N.
Kao5, J.M. Chen6, M. Abdel-Hafiez7, A.U.B. Wolter7, and R. Klingeler8
1
Low Temperature Physics and Superconductivity Department,a
M.V. Lomonosov Moscow State University, 119991 Moscow, Russia
2
Southern Federal University, 344090 Rostov-on-Don, Russia
3
Institute of Metal Physics, 620990 Ekaterinburg, Russia
4
Ural Federal University, 620002 Ekaterinburg, Russia
5
Institute of Physics, National Chiao-Tung University, 30076 Hsinchu, Taiwan
6
National Synchrotron Radiation Research Center, Hsinchu 30076, Taiwan
7
Leibniz Institute for Solid State and Materials Research, IFW Dresden, 01171 Dresden, Germany
8
Kirchhoff Institute for Physics, University of Heidelberg, 69120 Heidelberg,Germany
The doping series Eu1-xCaxCoO3-δ provides a rather peculiar way to study the spin-state
transition in cobalt-based complex oxides since partial substitution of Eu3+ ions by Ca2+
ions does not increase the mean valence state of cobalt but is accompanied by appearance
of oxygen vacancies in the ratio δ ~ х/2. In the parent compound EuCoO3, the low spin
(LS) – high spin (HS) transition takes place at temperatures so high that the chemical decomposition prevents its direct observation. The substitution of Eu3+ for Ca2+ in this system
shifts the LS-HS transition to lower temperatures. The energy gap ∆ associated with this
transition in octahedrally-coordinated Co3+ ions changes from 1940 K in EuCoO3 to 1540
K in Eu0.9Ca0.1CoO2.95 and 1050 K in Eu0.8Ca0.2CoO2.9. Besides, each O2- vacancy reduces
the local coordination of two neighboring Co3+ ions from octahedral to pyramidal thereby
locally creating magnetically active sites which couple into dimers. These dimers at low
temperatures form another gapped magnetic system with very different energy scale, D ~ 3
K, on the background of intrinsically non-magnetic lattice of octahedrally-coordinated
low-spin Co3+ ions.
Introduction
Among the transition metal oxides the cobalt-based ones exhibit the richest variety of valence and spin states.1 In RCoO3 perovskites (R is a rare-earth) the cobalt ions are presumed to
be in the 3+ state retaining 6 electrons in the d-shell.2 Depending on the ratio of local crystal
field splitting ∆CF between t2g and eg levels and the Hund’s coupling JH, the Co3+ ions may exhibit either low-spin (t2g6eg0), intermediate-spin (t2g5eg1), or high-spin (t2g4eg2) states. The crystal
field splitting is defined by the oxygen environment of the Co3+ ions. In case of a Hund’s energy
predominance and octahedral coordination, the high-spin state (S = 2) is Jahn-Teller active due
to partial filling of the t2g shell which yields local distortions of the ligand cages. At the increase
of ∆CF, the stable low-spin state (S = 0) can be realized. The possibility of the intermediate-spin
state (S = 1) is still under debate and might be derived from multiplet calculations when including spin-orbit coupling and the Jahn-Teller instability in both the t2g and eg segments of the dshell.3,4 Due to the subtle balance of ∆CF and JH in LaCoO3, the crossover between LS, IS and/or
HS states can be studied experimentally since modest variation of temperature induces changes
in the spin state between the low-spin state at low temperature and a magnetic IS or HS state. In
LaCoO3, the magnetization passes through a broad maximum at about 100 K and follows the
Curie law at elevated temperatures.5 At even higher temperatures, the magnetization increases
again and shows temperature independent behavior inherent for itinerant magnetic systems.6
1
This metal-insulator transition at about 500 K is usually ascribed to the melting of the orbitally
ordered state.
The substitution of La3+ ions by any other R3+ ions of smaller radius results in chemical
pressure which shifts both the spin state transition and the metal-insulator transition to higher
temperatures7-10. The effects are very different for both transitions, however. For instance, in
EuCoO3 the spin state transition takes place at about 2000 K while the temperature of the metalinsulator transition temperature is about 600 K. This strongly different effect of chemical pressure on both transitions emphasizes the different origin of the spin state and the metal-insulator
transitions.11
The substitution of R3+ ions by alkaline earth AE2+ ions may cause the appearance of tetravalent Co4+ ions. This should drastically change the magnetic and electronic properties in R13+
xAExCoO3 series of compounds due to presumed double exchange interactions between Co and
4+
Co ions. To be specific, ferromagnetism and increased conductivity may occur while, at commensurate doping levels, metal-insulator transitions can occur due to charge ordering phenomena.12,13 This scenario however demands exact oxygen stoichiometry in the material. The preparation of R1-xAExCoO3 compounds at ambient oxygen pressure usually results in oxygen deficient
samples R1-xAExCoO3-δ. In this case, the formation of Co4+ ions is energetically unfavorable and
only Co3+ ions are present. Since AE2+ doping is associated with chemical pressure significant
effects on the spin-gap and the metal insulator transition are to be expected. In addition, the removal of O2- ions from the crystal structure should change the local environment of both rareearth and transition metals in the oxygen deficient samples. While this effect is less relevant for
the rare-earths with their deeply hidden f-shells, the reduction in transition metal coordination
may change the spin state of electrons in the d-shells.
In the present work, X-ray diffraction (XRD) and X-ray absorption near edge structure
(XANES) spectra were taken on Eu1-xCaxCoO3-δ series of compounds in order to evaluate the
trends in the crystal lattice parameters and local environment of the transition metal Co upon
substitution of trivalent Eu for divalent Ca and oxygen deficiency. The specific heat and magnetization in Eu1-xCaxCoO3-δ were studied in order to reveal the associated effects on the thermodynamic properties at low temperatures and to evaluate trends in the spin-state transitions upon
doping at high temperatures. The choice of Eu3+ ions among the rare earths metals was motivated
by the fact that these ions exhibit a half-filled f-shell and hence a vanishing total magnetic moment, i.e. the rather small contribution of the transition metal subsystem is not masked. The Ca2+
ions among alkaline earths metals were chosen because the difference in ionic radii with the Eu3+
ions was expected to shift the spin-state transitions to lower temperatures.
The synthesis and XRD spectra
Polycrystalline samples of Eu1-xCaxCoO3-δ with 0 ≤ x ≤ 0.4 were prepared by the chemical homogenization (paper synthesis) method.14 At first, non-concentrated water solutions of
metal nitrates Eu(NO3)3, Ca(NO3)2 and Co(NO3)2 of 99.95 % purity were prepared. The exact
concentration of the dissolved chemicals was determined by gravimetry and in case of Co-based
solutions by means of complexonometric titration. The weighted amounts of metal nitrates solutions were mixed in stoichiometric ratios and calculated mixtures of nitrates were dropped on the
ash-free filters. The filters were dried at about 80°C and the procedure of the solutions dropping
was performed repeatedly. Then, the filters were burned and the remaining ash was thoroughly
grinded. It was annealed at 600°C for 4 hours to remove the carbon. The powder obtained was
pressed into pellets and sintered at 800°C in oxygen atmosphere for 30 hours. Finally, the samples were slowly cooled down to room temperature by switching off the furnace. The oxygen
content in the samples studied was determined by means of iodometric titration with 2% precision. The chemical analysis shows that at x > 0.2 the decomposition of the samples occurs, so
that for further investigations the samples with x = 0, 0.1 and 0.2 were selected.
The samples of Eu1-xCaxCoO3-δ (x = 0, 0.1, and 0.2) were analyzed at room temperature by
powder X-ray diffraction in symmetrical reflection geometry using Rigaku SmartLab (Cu Kα
2
radiation, λ=1.54187Å, graphite monochromator). All detectable peaks were indexed in the orthorhombic space group Pnma, as given in Table 1.
The full-profile Rietveld refinement was performed using JANA2006 package.15 The representative XRD pattern for EuCoO3 is shown in Fig. 1. The patterns were fitted with ten-term
Legendre polynomial background and four-term pseudo-Voight shape function. The zero-shift
and sample roughness were also refined. For all crystallographically independent atoms the
atomic positions and isotropic thermal parameters were refined. The oxygen thermal parameters
were constrained to be identical. The refinement of oxygen occupancies demonstrates the large
correlations between occupancies, thermal parameters and sample roughness. So, unfortunately,
it could not be refined against the present XRD data.
The structure of the Eu1-xCaxCoO3-δ (x = 0, 0.1, and 0.2) consists the eight-fold coordinated Eu (x=0) or Eu/Ca (x=0.1, 0.2) atoms and six-fold coordinated Co atoms in distorted octahedral environment. There are two non-equivalent positions for the oxygen ions O1 and O2
which constitute distorted CoO6 octahedron connected via shared corners. The four O2 atoms are
arranged in the equatorial plane while two O1 atoms are situated at the octahedron vertices. Rather small changes in the unit cell parameters, the unit cell volume remains practically constant,
reflect the fact that the withdrawal of the O2- ions is practically compensated by substitution of
Eu3+ ions by larger Ca2+ ions. As follows from the Table 2, the structural changes in the Eu1xCaxCoO3-δ (x = 0, 0.1, and 0.2) series of compounds are not univocal. The various Eu-O and
Co-O distances change monotonically with the increasing x but in different directions. Thus, the
average Eu-O and Co-O distances remain constant equal to 2.439Å and 1.930Å respectively.
Partial substitution of Eu by Ca leads to the increasing distortion of CoO6 octahedron.
Moreover the Co – Co distance shrinkage along the b axis, i.e. distance between Co ions bound
by O1 ions, with the simultaneous increase of Co – Co distance in the ac-plane. This could be
explained by withdrawal of apical oxygen ions O1 with the appearance of Ca2+ ions.
The removal of any oxygen changes the local environment of two neighboring Co3+ ions
from octahedron to pyramid. As will be shown below, these changes in local environment of the
Co3+ ions lead to changes in their spin state. To be specific, in pyramidal coordination the splitting of t2g and eg levels results in the formation of either the intermediate S = 1 state16 or the high
S = 2 spin-state.17 In both cases, magnetically active Co3+ ions arise which may interact and form
magnetic dimers.
The XANES spectra
The XANES spectra of Eu1-xCaxCoO3-δ (x = 0, 0.1 and 0.2) taken at liquid nitrogen temperature are shown in Fig. 2a. At partial substitution of Eu for Ca and accompanying oxygen
deficiency the positions of main L2,3 peaks remain virtually unchanged, but new features appear,
i.e. the pre-edge peak A strengthens, the C peak diminishes and strong peak D arises on the right
shoulders of both L2 and L3 edges. These extra D peaks at 784.5eV and 799 eV increase in amplitude with x in Eu1-xCaxCoO3-δ series of compounds.
Theoretical simulations of Co L3-edge XANES spectra in Eu1-xCoCaxO3-δ were carried
out using real-space Green’s function formalism implemented in self-consistent FEFF 8.4
code.18,19 Full multiple scattering approach with Hedin-Lundqvist exchange-correlation potential
was applied.20 Muffin-tin approximation was employed for the shape of potential. This method is
well-developed and has been proven to give useful structural information regarding substitutions,
vacancies and defects in various types of materials.21,22
The model used for calculations was the spherical fragment of EuCoO3 lattice (18 Å in
diameter), built around the absorbing cobalt atom. In this structure nearest neighbors of cobalt
are six oxygens occupying the O1 and O2 non-equivalent crystallographic sites. Although all O2
atoms are crystallographically equivalent, with respect to each particular cobalt atom there are
O2a and O2b atoms with slightly different Co-O distances, as shown in the inset to Fig. 2b.
Since doping of EuCoO3 by calcium causes the creation of oxygen vacancies, firstly the spectra
of pure and doped models with different number of oxygen vacancies were calculated to estimate
3
which factor (Eu-Ca substitution or creation of vacancies) has more influence on XANES spectra. In order to simulate the effect of doping, eight europium atoms, nearest to absorbing cobalt,
were replaced by calcium. It was found, that the only major consequence of Eu-Ca substitution is
the increase of pre-edge intensity (peak A). Apart from that, spectra of pure and doped material
with the same configuration of vacancies are very similar. It means that decrease of C feature
and appearance of D peak are mainly caused by the creation of oxygen vacancies, which occurs
as a result of the doping, but not by the Eu-Ca substitution itself.
Even the appearance of a single oxygen vacancy around the absorbing atom introduces
significant changes to the theoretical spectrum. Intensity of the C peak decreases with respect to
the spectrum without vacancies, but this decrease is not sufficient to turn the D shoulder into a
separate peak. Intensity and position of pre-edge peak A also differ from the experiment. To reproduce changes in the experimental spectrum multiple vacancy model was used. Final comparison between experimental and theoretical spectra for pure and Ca-doped EuCoO3 is shown in
Fig. 2b. Theoretical spectra are broadened in order to match the experimental resolution. Positions and relative intensities of peaks are in good agreement with the experiment for both pure
and doped samples. The only significant mismatch is too low intensity of D shoulder in the theoretical spectrum of doped model.
To check the origins of the present mismatch and the possible effect of chemicals contamination on these extra features, a specially designed experiment was conducted. XANES spectra were compared as shown in Fig. 3 for pure rare-earth cobaltates EuCoO3, SmCoO3, the Cadoped cobaltates Eu0.8Ca0.2CoO2.9 and Sm0.8Ca0.2CoO2.9, and the mechanical mixture of SmCoO3
and 0.4Ca(NO3)2 which contains significant excess of Ca as compared to Eu0.8Ca0.2CoO2.9 and
Sm0.8Ca0.2CoO2.9 samples. Indeed, the XANES spectra of SmCoO3 + 0.4Ca(NO3)2 sample contains additional features on the right shoulders of both L3 and L2 peaks, as shown in Fig. 3. Tentatively, this can be attributed to the presence of small quantities of Ba in the sample namely in
Ca(NO3)2, so that the barium M4,5 edge is visible in the experimental spectra and overlaps with D
feature. From the data in Fig. 3, nominal purity Ca(NO3)2 chemical provides about half of the
observed feature D intensity. Nevertheless, additional features seen on the right shoulders of both
L3 and L2 peaks are significantly more pronounced in magnitude. In theory the effect of Ba contamination is not taken into account; therefore, these calculated peaks are less pronounced compared with the observed ones.
Specific heat
The temperature dependences of specific heat C were measured in the range 2 – 30 K in
“Quantum Design” Physical Properties Measurements System PPMS-9T by the relaxation technique. The C(T) curves obtained for Eu1-xCaxCoO3-δ series of compounds are shown in Fig. 4. In
this temperature range the C(T) dependence in the parent compound EuCoO3 can be well approximated by just the cubic term βT3 but a tiny Schottky – type anomaly is observed at low
temperatures, nevertheless. This anomaly becomes significantly more pronounced in the
Eu0.9Ca0.1CoO2.95 and Eu0.8Ca0.2CoO2.90 samples. Therefore, at low temperatures the specific heat
can be approximated by the sum C = βT3 + CSch.
As follows from XRD and XANES analyses, the partial substitution of trivalent Eu by
divalent Ca is accompanied by the appearance of apical oxygen vacancies. Taking into account
that the removal of O2- ions reduces the coordination number of the neighboring Co3+ ions, one
can assume that the position of cobalt with respect to oxygens will change in a way that the cation ion will move within the pyramid. In this case the spin state of Co3+ may change from the
non-magnetic LS (S = 0) to a magnetic either intermediate IS (S = 1) or high spin HS (S = 2)
state. Two neighboring ions in pyramid coordination may weakly interact with each other forming the magnetic dimers. Assuming the magnetic nature of the Schottky – type anomaly one can
approximate its shape by the formula for S = 1
4
 J 
C = 6nR 
 kT 
2
 2 J 
 4J 
 6J  
exp −
1 + 15 exp −
 + 20 exp −
 
kT
kT




 kT  

 2J 
 6J  
1 + 3 exp −
 + 5 exp −
 
 kT 
 kT  

(1)
2
or similar but more cumbersome expression for S = 2, where n is the concentration of dimer entities, R = 8.314 J/mol K is the universal gas constant, and J is the exchange interaction parameter
within the dimer. The fits of available data by Eq. 1 are shown in the Inset to Fig. 4. Within the
limits of accuracy, fitting by means of Eq. 2 describes the data within similar quality, i.e. the data
do not discriminate between the S=2 and the S=1 case. The estimation of the phonons contribution gives β = (1.7±0.1)×10-4 which corresponds to the Debye temperature of about 400 K. The
gap value deduced from the Schottky anomaly position was found to be D = 2J = 3.2 ± 0.1 K,
while for Eu0.9Ca0.1CoO3-δ and Eu0.8Ca0.2CoO3-δ samples the concentrations n were found to be
0.052 and 0.098, correspondingly. These numbers (n = x/2) are in good agreement with expected
0.05 and 0.1 for x = 0.1 and 0.2.
Magnetization
The temperature dependences of the magnetization in the Eu1-xCaxCoO3-δ family of compounds were measured by means of a VSM-SQUID magnetometer (Quantum Design) in the
temperature range 2 – 390 K in both zero-field-cooled (ZFC) and field-cooled (FC) regimes at B
= 0.1 T. In case of the parent compound EuCoO3, ZFC and FC curves practically coincide in the
whole temperature range measured. Therefore, hereinafter the curves measured in ZFC regime
are discussed only.
The temperature dependences of the static magnetic susceptibility χ = M/B for the family
of Eu1-xCaxCoO3-δ samples taken at B = 0.1 T are shown in Fig. 5. Firstly, the magnetic response
of the undoped material EuCoO3 is discussed. The data are in a good qualitative agreement to
previous studies on single- and polycrystalline materials.11 As shown in Fig. 5, below about T =
250 K, the magnetization is nearly perfectly described by the van Vleck magnetism of the Eu3+
ions and a Curie-Weiss-like divergence of the magnetic susceptibility. Temperature independent
diamagnetic and Co3+ van Vleck contributions are very small in comparison. The Curie-Weisslike contribution implies a negligible number of paramagnetic sites in the material, i.e. of either
HS or IS Co-ions or Eu2+ ions, cf. with tiny Schottky-type anomaly found in the specific heat of
the nominally pure parent compound EuCoO3. At higher temperature there is an additional contribution to the magnetization which we ascribe to the temperature-induced evolution of Co3+
ions in the HS or IS state. A quantitative analysis of this feature will be given below.
Upon Ca-doping, several effects show up in the data in Fig. 5: (i) there is a strong increase of the Curie-Weiss-like low-temperature response; (ii) the pronounced shoulder at intermediate temperature which signals a dominant Eu3+ van Vleck contribution to the susceptibility
in this temperature range is suppressed; (iii) the susceptibility at high temperatures increases.
Qualitatively, these features suggest the increase of paramagnetic sites at low-temperatures, the
decreasing relevance of the Eu3+ van Vleck magnetism and the decrease of the magnetic excitation gap associated with the Co3+ HS- or IS-state formation.
In the following, the various contributions to the susceptibility will be discussed in more
details. In accordance with the analysis of our XRD and XANES data, we assume that the substitution of Eu3+ ions by Ca2+ ions is accompanied by the formation of Co3+ ion pairs in pyramidal
coordination, i.e. of pairs of paramagnetic Co3+ sites forming antiferromagnetically coupled
magnetic dimers. Generally, the expression for the magnetic susceptibility in the Eu1-xCaxCoO3-δ
family of compounds can hence be written as a sum of virtually independent terms:
χ = χ0 + χEu + χCo-pyr + χСо-oct
(2)
5
where χ0 is the temperature-independent diamagnetic and Co van Vleck contribution, χEu is the
van Vleck contribution of the Eu3+ ions, χCo-pyr is the contribution of magnetically active Co3+
ions in pyramidal coordination, and χСо-oct is the Со3+ ions contribution dependent on their spin
state in octahedral coordination.
The procedure of separation of various terms to the magnetic susceptibility could be illustrated for the parent compound EuCoO3, as follows. As discussed above, at low temperatures,
the main contributions to the magnetization are due to the Eu3+ van Vleck term χEu and a CurieWeiss-like term due to a small amount of paramagnetic sites. The van Vleck term can be written
as
24
) + (13.5 − 1.5γ )e −γ + (67.5 − 2.5γ )e −3γ + ...
2 (
N µ
γ
χ Eu = A B
(3)
3k B T
1 + 3e −γ + 5e −3γ + ...
The only fitting parameter in this formula γ = λ/kBT is a constant of the spin-orbital interaction, i.e. the multiplet width.23 This contribution is almost temperature independent below ca.
100 K and decreases proportional to 1/T at high temperatures. The best fit of the experimental
data at intermediate temperatures could be obtained with the multiplet width λ = 311 cm-1 (~ 450
K). This value is in good correspondence with the values of this parameter in other Eu-based
oxides.24,25 In Figs. 5 and 6, the results of fitting of the van Vleck contribution χEu are shown as
dashed lines. Note, that the summation of the Pascal’s constants for EuCoO3, i.e. χ0 = - 6.6×10-5
emu/mol,26 and a Co3+ van Vleck contribution are negligible as compared to other terms.
As mentioned above, our data suggest that all relevant contributions to the magnetic susceptibility change for the samples with x ≠ 0. In such a situation, when the magnetic susceptibility χ comprises contributions differently dependent on temperature, it is more convenient to
represent the experimental data as temperature dependences of the product χ·Т, as shown in Fig.
6. In this case, hyperbolic terms will be temperature independent.
In the oxygen-deficient samples Eu1-xCaxCoO3-δ, i.e. for x = 0.1 and 0.2, the quantity
χ·Т demonstrates doping-induced changes both at low and high temperatures while at intermediate temperatures there are only small effects observed. However, as will be shown in more
details below we ascribe the latter to an incidental compensation of decreasing Eu3+ van Vleck
magnetism by the increasing relevance of the other terms.
At low temperature, upon doping a clear maximum evolves in χ·Т. Indeed, such a feature can be straightforwardly explained in the frame of our analysis by the formation of magnetic
dimers. In this scenario, the magnetic susceptibility probes the dimer excitation gap, i.e. the size
of the gap is given by the peak maximum, the associated susceptibility changes measure the
number of dimers, and the peak width provides information on the distribution of gap energies.
Our data hence indicate that the number of dimers increases upon Ca-doping and that the magnetic interaction within the dimers is only small, i.e. of the order of few Kelvin degrees, which
agrees both with the specific heat measurements and the band structure calculation results presented in the next section.
At high temperatures, the divergence of curves, shown in Fig. 6, is due the contribution
3+
of Co ions in octahedral coordination. At elevated temperatures, these Co3+ ions gradually
transform from the non-magnetic low spin state (S = 0) to a magnetic either intermediate spin (S
= 1) or high spin (S = 2) state. The temperature dependence of the magnetic susceptibility χСо-oct
can be hence approximated by the expression for two-level system:11
N g 2 µ B2 S ( S + 1) ν (2S + 1)e − ∆ / T
χ (T ) = A
(4)
3k B T
1 + ν (2S + 1)e − ∆ / T
where ν = 6 is the degeneracy of the intermediate spin state and ν = 3 is the degeneracy of the
high spin state in Co3+ ions, the gap ∆ is the splitting between the low-spin LS and excited either
6
IS or HS state. An estimation of the gap ∆ can be made directly from this expression by analyzing the high temperature magnetic susceptibility.
In Fig. 6 the product χ·Т in the parent compound EuCoO3 is compared to the Eu3+ van
Vleck contribution. The difference of these curves, shown in the lower inset of Fig. 6, can be
approximated by the magnetic susceptibility given by Eq. (5) multiplied by temperature. Fitting
the data yields the energy gap ∆ between the low-spin and intermediate-spin state in EuCoO3 of
∆ = 1940 K. This value nicely agrees to a previous analysis of single crystal data which yielded
∆ > 1900 K.11 Upon doping, the magnetization at high temperatures increases which already qualitatively indicates that the energy gap ∆ becomes smaller. Indeed, calculations similar to that for
the parent compound allow establishing the concentration dependence of the energy gap ∆ in the
Eu1-xCaxCoO3-δ family of compounds, as shown in Fig. 7. Evidently, the gap decreases rapidly in
the range doping 0 < x < 0.2 reaching ∆ = 1540 K in Eu0.9Ca0.1CoO2.95 and ∆ = 1050 K in
Eu0.8Ca0.2CoO2.9. It extrapolates through polynomial to zero at about x = 0.365.
Our data hence show that the energy gap ∆ between the low-spin and intermediate-spin
state strongly decreases upon doping. This behavior agrees to what is expected for changing the
chemical pressure in EuCoO3 by partial substitution of the europium ions (the ionic radius of
Eu3+ is 120 pm) for larger calcium ions (the ionic radius of Ca3+ is 126 pm). A similar tendency
is observed when Eu is replaced by smaller La as well as in hydrostatic external pressure experiments on LaCoO3.27
Intra-dimer magnetic exchange
In order to estimate the size of the spin gap in the dimers we performed calculations of
the exchange parameters J in the Heisenberg model within the density functional theory using
Green's function formalism.28 We used the linearized muffin-tin orbitals (LMTO) method29 and
the LDA+U approximation.30 The on-site Coulomb repulsion U and intra-atomic exchange parameter JH were chosen to be 7.8 eV and 0.99 eV respectively.5 The Eu-4f states were considered
as pseudocore.31 The integration in the Brillouin zone was performed over 144 k-points.
We started with the EuCoO3 crystal structure and constructed the supercell with 8 Co
atoms. According to our XRD data apical oxygen ions are primary removed with the doping. In
the calculations we followed the same scenario and removed these oxygens to simulate the dimers (formed by two pyramidal Co) directed along c-axis. The lack of the charge due to the oxygen withdrawal was compensated by the corresponding shift of the Fermi level. Since our aim is
to calculate the magnetic properties at lowest temperatures we assumed that all octahedral Co
ions are in the low-spin state, while pyramidal Co were allowed to be magnetic.
The exchange parameter J within the dimer was found to be 3.1 K. This is in agreement
with the specific heat and magnetic susceptibility measurements. Since in the high-spin state
stabilized in the pyramidal Co3+ ions16 eg orbitals are half-filled they participate in the antiferromagnetic superexchange interaction within the dimer. In the present case, when the dimers are
along c-axis, Co3+ z2-orbitals are active. However, since in the high spin state both eg orbitals are
actually half-filled the same mechanism (but with x2-y2 orbital) is expected for the planar order
of the dimers Co, even if such will be the case.
Conclusion
The synthesis of Eu1-xCaxCoO3-δ at ambient pressure yields a series of compounds with
unchanged Co valence states since the partial substitution of Eu3+ ions for Ca2+ ions is accompanied by oxygen deficiencies, as observed in XRD and XANES measurements. This family of
compounds hence provides an alternative way to change the local structure and to study the spin
state transitions. Here, we have presented XANES, XRD, specific heat, and magnetization data
in order to elucidate the various contributions to the magnetic response in this system where only
Co3+ ions are present. In octahedral coordination, these ions exhibit a non-magnetic S = 0 ground
state at low temperatures and experience a spin state transition at increasing temperatures. The
7
energy gap ∆ associated with LS-HS transition in octahedrally-coordinated Co3+ ions changes
from 1940 K in EuCoO3 to 1540 K in Eu0.9Ca0.1CoO2.95 and 1050 K in Eu0.8Ca0.2CoO2.9. The
analysis of the product χ·Т allows revealing the major trends in behavior of the systems experiencing the transition between different spin states.
The removal of the oxygen ion from the perovskite crystal lattice changes the local environment of two neighboring cobalt ions from octahedral to pyramidal. The Co3+ ions in pyramidal coordination are in a magnetic (either S = 1 or S = 2) ground state at low temperatures. The
weak interaction between these spins results in the formation of magnetic dimers which show up
in the low temperature magnetic susceptibility χ and specific heat C through appearance of
Schottky – type anomaly with D ~ 3 K. The intra-dimer exchange calculation performed within
the density functional theory confirms this estimation.
Acknowledgements
We thank K.I. Kugel, D.I. Khomskii, B. Büchner, and Z. Hu for fruitful discussions. This
work was supported through RFBR Grant No. 10-02-00021, 10-02-96011, 11-02-91335 and by
the DFG via 486 RUS 113/982/0-1.
References
1. M. Imada, A. Fujimori, Y. Tokura, Rev. Mod. Phys. 70, 1039 (1998).
2. S. Yamaguchi, Y. Okimoto, Y. Tokura, Phys. Rev. B 54, 11022 (1996).
3. A. Ishikawa, J. Nohara, S. Sugai, Phys. Rev. Lett. 93, 136401 (2004).
4. R. S. English, J. Wu, and C. Leighton, Phys. Rev. B 65, 220407 (2002).
5. M.A. Korotin, S.Y. Ezhov, I.V. Solovyev, V.I. Anisimov, D.I. Khomskii, G.A. Sawatzky, Phys. Rev. B 54, 5309 (1996).
6. C. Zobel, M. Kriener, D.Bruns, J. Baier, M. Gruninger, T. Lorenz, P. Reutler, A. Revcolevschi, Phys. Rev. B 66, 020402 (2002).
7. S. Yamaguchi, Y. Okimoto, Y. Tokura, Phys. Rev. B, 54, 11022 (1996).
8. J.-S. Zhou, J.-Q. Yan, J. B. Goodenough, Phys. Rev. B 71, 220103 (2005).
9. M. Tachibana, T. Yoshida, H. Kawaji, T. Atake, E. Takayama-Muromachi, Phys. Rev.
B. 77, 094402 (2008).
10. K. Berggold, M. Kriener, P. Becker, M. Benomar, M. Reuther, C. Zobel, T. Lorenz,
Phys. Rev. B 78, 134402 (2008).
11. J. Baier, S. Jodlauk, M. Kriener, A. Reichl, C. Zobel, H. Kierspel, A. Freimuth, T. Lorenz, Phys. Rev. B 71, 014443 (2005).
12. D. Louca, J. L. Sarrao, J. D. Thompson, H. Roder, G. H. Kwei, Phys. Rev. B 60, 10378
(1999).
13. D. Louca, J. L. Sarrao, Phys. Rev. Lett. 91, 155501 (2003).
14. S. Quezel-Ambrunaz, Bull. Soc. Fr. Mineral. Crystallogr. 91, 339 (1968).
15. V. Petricek, M. Dusek and L. Palatinus, Jana2006. Structure Determination Software
Programs. Institute of Physics, Praha, Czech Republic (2006)..
16. C. Frontera, J.L. Garcıa-Munoz, A. Llobet, M.A.G. Aranda, Phys. Rev. B 65, 180405
(2002).
17. Z. Hu, H. Wu, M. W. Haverkort, H. H. Hsieh, H.-J. Lin, T. Lorenz, J. Baier, A. Reichl,
I. Boon, C. Felser, A. Tanaka, C. T. Chen, L. H. Tjeng, Phys. Rev. Lett. 92, 207402
(2004).
18. J.J. Rehr and R.C. Albers, Rev. Mod. Phys. 72, 621 (2000).
19. J.J. Rehr and A.L. Ankudinov, Coordination Chem. Rev. 249, 131 (2005).
20. L. Hedin and B.I. Lundqvist, Journal of Physics C: Solid State Phys. 4, 2064 (1971).
21. A. Soldatov, D. Lamoen, M. Konstantinovic, S. Van den Berghe, A. Scheinost, M.
Verwerft, J. Solid State Chem. 180, 54 (2007).
22. Q. Liu, W. Yan, H. Wei, Z. Sun, Z. Pan, A. V. Soldatov, C. Mai, C. Pei, X. Zhang, Y.
Jiang and S. Wei, Phys. Rev. B 77, 245211 (2008).
8
23. K.Tezuka, Y. Hinatsu, N.M. Masaki, M. Saeki, J. Solid State Chem. 138, 342 (1998).
24. Y. Takikawa, S. Ebisu, S. Nagata, J. Phys. Chem. Solids 71, 1592 (2010).
25. G.F. Goya, R.C. Mercader, M.T. Causa, M. Tovar, J. Phys.: Condens. Matter 8, 8607
(1996).
26. G.A. Bain and J.F. Berry, J. Chem. Educ. 85, 532 (2008).
27. K. Asai, O. Yokokura, M. Suzuki, T. Naka, T. Matsumoto, H. Takahashi, N. Môri, K.
Kohn. J. Phys. Soc. Japan 66, 967 (1997).
28. A.I. Liechtenstein, V.A. Gubanov, M.I. Katsnelson, V.I. Anisimov, J. Magn. Magn.
Mat. 36, 125 (1983).
29. O. K. Andersen and O. Jepsen, Phys. Rev. Lett. 53, 2571 (1984).
30. V.I. Anisimov, F. Aryasetiawan and A.I. Lichtenstein, J. Phys.: Condens. Matter 9, 767
(1997)
31. I. A. Nekrasov, S. V. Streltsov, M. A. Korotin and V. I. Anisimov, Phys. Rev. B 68,
235113 (2003).
9
Fig. 1. Room temperature powder XRD pattern (black circles) of EuCoO3 in the 2θ
range 10 – 90°, Rietveld refinement fit (solid red line), difference profile (lower solid green line),
and positions of Bragg peaks (vertical bars). In the inset the primitive cell of EuCoO3 is shown.
The Eu ions are indicated by large gray circles, the Co ions both bare and in octahedral oxygen
environment are represented by blue circles and two positions of oxygen ions are shown by small
red (O1) and black (O2) circles.
Fig. 2. The Co L2,3-edge XANES spectra in Eu1-xCaxCoO3-δ series of compounds taken
at liquid nitrogen temperature (a); the comparison of experimental and simulated XANES spectra in pure EuCoO3 and Ca-doped compounds (b). The spectra were shifted along y-axis for clarity.
Fig. 3. The Co L2,3-edge XANES spectra in EuCoO3, SmCoO3, Eu0.8Ca0.2CoO2.9
Sm0.8Ca0.2CoO2.9 and in the mechanical mixture of SmCoO3 and 0.4Ca(NO3)2 powders.
Fig. 4. The temperature dependences of specific heat C in Eu1-xCaxCoO3-δ series ofcompounds. Inset: the approximation of the specific heat in Ca-doped compounds by sum of
cubic term βT3 (not shown) and Schottky-type anomaly CSch (dashed lines).
Fig. 5. Temperature dependence of the magnetic susceptibility χ of Eu1-xCaxCoO3-δ.
The dashed line represents the Eu3+ van Vleck contribution for the parent compound EuCoO3.
Fig. 6. The temperature dependences of the product χ·Т for the Eu1-xCaxCoO3-δ family
of compounds. Lower inset represents the difference between the data for x = 0 and the Eu3+ van
Vleck contribution (black data and dashed line, respectively, in main plot). The upper inset highlights the low temperature behavior.
Fig. 7. The concentration dependence of the gap ∆ between low spin and high spin
states in the Eu1-xCaxCoO3-δ family of compounds. The dashed line is a polynomial extrapolation
of the experimental data.
10
Fig. 1.
Eu0.9Ca0.1CoO2.95
Co L3
Co L2
EuCoO3
770
780
790
B
b
Eu0.8Ca0.2CoO2.90
Normalized Absorption (arb. units)
Normalized Absorption (arb. units)
a
Co L3
C
A
D
Exp. pure
Theory pure
Exp. doped
Theory doped
800
775
780
785
790
Photon Energy (eV)
Photon Energy (eV)
Fig.2.
11
SmCoO3
1.0
Normalized absorption (arb. units)
EuCoO3
Sm0.8Ca0.2CoO2.9
0.8
Eu0.8Ca0.2CoO2.9
SmCoO3+0.4Ca(NO3)2
0.6
0.4
0.2
0.0
775
780
785
790
795
800
Photon Energy (eV)
Fig. 3
5
Eu0.9Ca0.1CoO3-δ
Eu0.8Ca0.2CoO3-δ
C (J/mol*K)
4
3
2
4
6
8
EuCoO3
T (K)
1
Eu0.9Ca0.1CoO2.95
Eu0.8Ca0.2CoO2.90
0
0
5
10
15
20
25
30
T (K)
Fig. 4.
12
8 0
E u
1 -x
C a
x
x
x
-3
χ (10 emu/mol)
6 0
x
=
=
=
χ
E
V
C o O
8
3 -δ
0
0 .1
0 .2
u 3 +
(x = 0 )
V
7
4 0
6
2 0
5
0
0
1 0 0
2 0 0
T (K )
3 0 0
4
4 0 0
Fig. 5.
2.0
0.4
x=0
x=0.1
x=0.2
0.2
Eu3+
χVV
0.0
0
10
20
30
40
0.12
1.0
Eu3+
[χ(x=0)-χVV )]⋅T
χ⋅T (emu⋅K/mol)
1.5
0.5
0.10
0.08
0.06
0.0
0
100
200
300
400
T (K)
Fig. 6.
13
2000
1800
1600
1400
∆ (K)
1200
1000
800
600
x = 0.365
400
200
0
0.0
0.1
0.2
0.3
0.4
x
Fig. 7.
14
Table 1. Crystallographic parameters of the orthorhombic Eu1-xCaxCoO3- system at room temperature (space grooup
Pnma). Listed are the lattice parameters a, b and c and the quality-of-fit parameters R1/ωR2, Rp1/ωRp2, and Goodness-ofFit (GOF).
a (Å)
b (Å)
c (Å)
V, Å3
Nobs/Ntotal/
Nstr.param/
Ntotal param
R1/
ωR2
Rp/
ωRp
GOF
EuCoO3
5.3715(2)
7.4785(2)
5.2572(2)
211.19(2)
90/94/10/28
0.0160/
0.0202
0.0843/
0.1256
1.69
Eu0.9Ca0.1CoO2.95
5.3653(2)
7.4763(2)
5.2632(2)
211.12(2)
92/94/9/28
0.0147/
0.0192
0.0690/
0.1074
1.58
Eu0.8Ca0.2CoO2.90
5.3609(2)
7.4749(3)
5.2691(2)
211.15(2)
91/93/9/26
0.0166/
0.0231
0.0710/
0.1103
1.72
Compound
Table 2. Selected interatomic distances in Eu1-xCaxCoO3- series of compounds at room temperature.
Parameter
EuCoO3
Eu0.9Ca0.1CoO2.95
Eu0.8Ca0.2CoO2.90
Eu1 – O2iii, Eu1 – O2iv
2.320(9)
2.314(8)
2.311(11)
Eu1-O2, Eu1-O2v
2.506(9)
2.516(8)
2.528(11)
Eu1-O2i, Eu1-O2vi
2.605(9)
2.604(8)
2.600(11)
Eu1 – O1i
2.294(11)
2.285(11)
2.264(14)
Eu1 – O1ii
2.359(11)
2.364(10)
2.374(12)
Co1-O2, Co1-O2vii
1.926(9)
1.922(8)
1.912(11)
Co1-O2i, Co1-O2viii
1.942(9)
1.947(8)
1.956(10)
Co1-O1, Co1-O1ii
1.921(3)
1.921(2)
1.925(3)
Co-Coi (in ac plane)
3.7580(2)
3.7579(2)
3.7584(1)
Co-Coii (along b axis)
3.7393(2)
3.7381(2)
3.7374(2)
Symmetry codes: (i) −x+1/2, −y, z+1/2; (ii) −x, y+1/2, −z; (iii) x+1/2, −y+1/2, −z−1/2; (iv) x+1/2, y, −z−1/2; (v) x,
−y+1/2, z; (vi) −x+1/2, y+1/2, z+1/2; (vii) −x, −y, −z; (viii) x−1/2, y, −z−1/2.
15